The Journal of Strain Analysis for Engineering Design

An examination is made of a number of neutral loading experiments according to classical plasticity theory. For work hardening materials the question as to whether plastic deformation occurs during neutral loading depends strongly upon the deformation produced from initial loading. Initial elastic loading with a subsequent stress path that follows the boundary of the initial vield surface is truly neutral with a wholly elastic response. However, when plastic strain is produced from initial loading then plastic flow is produced from a subsequent stress path that follows the boundary of a surface that is an isotropic expansion of the initial yield surface. Since this violates the assumption of isotropic hardening the usefulness and limitations of the rule of kinematic hardening are examined. It is further shown that the results from recent experiments are particularly relevant to the appraisal of modern developments to plasticity theory.

Neutral loading without work hardening will produce plastic flow. It is shown that the corresponding Prandtl—Reuss theoretical solutions are representative of observed behaviour for both severely pre-strained materials and for others that do not harden appreciably.

The concept is introduced of lower and upper bounds on the deformation that can be expected from neutral loading. These correspond to the extreme purely elastic and non-work-hardening solutions, respectively.

A practical, compact mixed-mode (CMM) fracture specimen, which can be used for any combination of mode-I and mode-II stress intensities, is described. The loading system, which is suitable for high temperature testing, is also described. The specimen can be used for mixed-mode testing under static, fatigue, creep, or creep/fatigue conditions.

The stress-intensity factors (K_I and K_II) for the complete range of loading conditions have been obtained by finite element and photoelastic methods. Nodal displacement, stress, and contour integral methods, using the finite element results all gave practically the same stress intensity factors. These were in reasonably close agreement with the results obtained using shear stress distributions obtained from the photoelastic tests.

This paper investigates the contact problem in which an elastic strip is indented by a rigid body (punch) of arbitrary shape. Both bonded and unbonded strips are considered.

A numerical method due to Gladwell (1)^† is shown to be a direct and effective technique for analysing the effect of any punch whose profile is a polynomial of degree n, over a range of a/t (semi-contact width to a depth ratio) which is of practical interest 0 ≤ a/t ≤ 10 for Poisson's ratio 0 ≤v ≤ 0.5.

For the cylindrical punch results are presented and compared with Meijers' asymptotic analytic solutions (2). For small a/t agreement is very good as expected. For a/t large, however, there are some large discrepancies which can be traced to an error in Meijers— expression for pressure distribution when v ≠ 0.5. Results are also presented for both the flat and the linear punch.

Tensile tests were performed on a straight steel strand of three layer (12/6/1) construction, having a core wire diameter of 3.66 mm and 3.33 mm diameter helical wires, under conditions of full end-fixity, partial restraint, and ends free from torsional restraint. The torque generated under tensile load was recorded as well as the strand extension and rotation over a 600 mm gauge length. Wire tensions and bending moments in the outer layer of helical wires were determined at the mid-strand position from the outputs of strain gauges in groups of three with parallel grids and mounted parallel to the wire axis on the crown of each wire.

The rate of strand extension under tensile load was found to be greater in tests with reduced torsional restraint, the greatest rate occurring in the free-end test. The strand rotation rate was also found to be greatest in the free-end test. The greatest difference from the theoretically predicted rates occurred in a free-end test with increasing load; predicted values of extension and rotation underestimated the test results by 12 and 23 per cent, respectively. Displacement of the load-torque plots occurred in the direction of reducing torque as testing proceeded. This appears to indicate the redistribution of the strand load between the layers of wires.

Wire tensions showed a more even sharing of load in the fixed-end condition than in the free-end condition. The increase in rate of tension with strand load was less for most wires in tests with reduced torsional restraint, with the lowest tension rates developing in the free-end condition.

For most wires, the rate of bending moment change with strand load was greater (in the sense tending to decrease tensile stress on wire crowns) in tests with reduced torsional restraint. However, the bending moment rates varied greatly between wires, the variation being greater in tests with reduced torsional restraint than in fixed-end tests.

Creep crack growth rates in the 32%−Ni−20%−Cr alloy Incoloy 800 H at 800°C are correlated with the fracture mechanics parameter C∗ integral. This was done by experimental and numerical investigations of different specimen sizes and geometries under constant load, constant rate of crack opening displacement or crosshead displacement, or slow cyclic loading. In the numerical simulations plane and three-dimensional finite element calculations have been performed without consideration of creep crack growth.

Carefully conducted experimental studies of the torsional characteristics of substantial structural strands are reported. The match between experimental results for an old and fully bedded-in 39 mm diameter, 91 wire spiral strand and some theoretical predictions is very encouraging. The theory, which has been reported elsewhere, treats the individual layers of wires in a strand as orthotropic sheets and, via established results in contact stress theory, takes full account of the frictional interactions between wires.

Static and dynamic torsional stiffness and hysteresis data are presented for axially-preloaded strands. Strain gauging of individual wires showed significant scatter, but the match between theory and the mean of the experimental data for a position remote from the termination was good, using the same assumed value of the friction coefficient for all tests.

For an old strand, random loading can significantly increase the level of torsional hysteresis above that found for regular cyclic loading, provided that the interwire force changes are large enough to overcome the contact friction.

Tests on a new 41 mm diameter strand indicated that hysteresis measurements on newly-manufactured strands could be misleading for long term applications.